jump to navigation

Useful Web Analytics Definitions November 14, 2012

Posted by Joe Kamenar in web analytics.

Here are some key terms and concepts that you will encounter as a web analyst. Depending on the type of industry you are in, you may find some of these useful in your analysis.

Metrics vs. Dimensions

In web analytics, you will often hear the terms “metrics” and “dimensions”, and it is important that you understand the difference between them. A metric is either a count or a ratio. A count can include visits, visitors, instances, time on site, page views, downloads, single-page visits, number of site searches,  orders, and other items that stand on their own. A ratio is a combination of two counts, such as pages per visit, conversions divided by starts (conversion rate), average order per visit, bounce rate and other combinations of two metrics.

Dimensions are items that you are measuring, such as browser type, location, pages, content groups, search terms, downloads, videos, and such. For each dimension, you apply one or more metrics. When you hear the term “drill down”, or “correlation”, you are essentially taking one dimension, and then looking at how a second dimension breaks down on the first dimension. For example, suppose you have two dimensions of interest – content groups and visitor types. You can do a one dimension report where you look at visits and page views to each content group, or you can do a two-dimension drill-down or correlation report where for each content group, you look at what user types engaged with that content group, based on the visit and page view metrics.


Logins vs. Visits

I see a lot of confusion in this area. Many business users assume that there is no difference between a login and a visit on a site like a portal, where a user must log in. They don’t understand why the two values can differ so greatly. Let me explain the reason for this confusion. A visit, in analytics-speak, is based on hits to tagged content on the website where no more than 30 minutes have elapsed between hits. If I am logged in to a portal, visit some pages or conduct some business, but then walk away from the site for 35 minutes, then come back and go to another page on the portal, I will be counted as a new visit. However, if the login security did not time me out from being logged in, I will not register a new login. Hence, we have one login, but two visits.

Conversely, I can come to a site, login, spend a few minutes doing my thing on the site, then log out. 10 minutes later, either me or someone else at my company logs in from the same computer. Now, I have two logins, but only one visit, as far as the analytics tool sees it. Keep in mind that there is no correlation between logins and visits, as each metric is tracked differently! It is rare that a company will build out their login sessions to match a web analytic session.

Key Performance Indicators (KPIs)

A KPI is essentially a key metric. Depending on your business, some key metrics can be “bounce rate”, “average order size”, “conversion rate”, cost per customer, “loyalty rate”, “banner click rate”, “self-service” rate, and others that give you an idea of how your site is meeting its objectives. For a pharma company I did some work for, their key KPIs focused on what percent of their visitors were “engaged”, “educated”, and “motivated”.  KPIs should reflect the key business goals of the site, and not things like visits, page views, pages per visit, visitors, and other basic metrics. KPIs need to be tied directly to site objectives, and be accepted by the organization.


Unique Visitors

A unique visitor is counted only once during a measurement period. This means that if I visit a site three times during a month, I will register as three separate visits, but only as one unique visitor. A lot of organizations obsess with the number of unique visitors. There is one big problem with the unique visitor metric – it is not accurate. A unique visitor is measured not always by sticking a cookie in a browser, but also by looking at the user-agent string. This is a combination that includes the browser type and visitor IP. Many business internet users (and some home internet users) do not have a “static” IP. They have what is known as a “dynamic” IP address, which means it can change during the week or month, as the internet provider assigns a new IP address from their pool. So, to your web analytics package, any single visitor can look like a different visitor each time they access the site. Thus, they are considered as “unique”, even if they are one and the same person, sitting at the same PC or laptop.


New vs. Return Visitors

This is typically a key visitor segment that you see in web analytic reports. It shows the breakout of new visitors vs. return visitors. A new visitor is a unique individual who made his first visit to your site during a reporting period. A return visitor is a unique individual who made a visit to the site during the current reporting and during a prior reporting period. Basically, a new visitor is someone who has never been to your site before, while a return visitor is someone who has been to the site over at least two reporting periods.


Bounce Rate

When a visitor comes to your website and leaves after viewing just that one page, this is known as a “bounce”. A key metric, known as the bounce rate, measures the percentage of time that a visitor bounces from any given page A high bounce rate is of concern when the page being visited comes from any sort of paid marketing, as the visitor clearly is not interested in going deeper into your site. As a web analyst, you need to identify your site’s landing pages and track the bounce rate over time for each page. It is critical to keep the bounce rate low for landing pages for paid media sources, as everyone who bounces is a wasted marketing spend, especially if the cost is based on a pay-per click model.


Time on Site

Many companies like to see “time on site”, or “TOS” as a key metric. I am not sold on this one. A visitor who comes to your site, looks at a page, then takes a call or responds to a text, while still on your site is not really spending that time reading content on your site or downloading key materials. What is important instead of the actual time on site metric is how this metric trends over time. If your average TOS starts to decrease, it can mean different things based on the purpose of your site. If in the past, a visitor had to spend a lot of time trying to find what they were looking for, but now, due to an improved FAQ or site search they find it quickly, the TOS may actually decline. Is that bad? Not necessarily. If your success metrics are showing an increase, then the decline in the TOS shows that your visitors are not wasting time looking for things.



When a visitor sees an ad or banner for your site, that is known as an “impression”.  When the person clicks on the link to get to the site, that is known as a click-through. The click-through rate (CTR) is a key metric for advertisers, as this is the percent of impressions that get clicked on.


Page Engagement Rate

A page engagement rate shows you how many times the page is viewed per visit. For example, if you have a page that had 1,000 visitors and that was viewed 1,500 times, then the page engagement rate is 1.5. Typically, topical stories or news articles may have an engagement rate close to 1.0, while pages such as your FAQ and other core pages may have an engagement rate that is higher.


Banner Fatigue

Banner fatigue is not a specific metric. Rather, it is a trend number that you need to monitor. If your site caters to mostly a fixed number of users, such as a business portal or physicians who are involved in a particular specialty, you will find that over time, users will stop clicking on banners on the site. These banners are typically internal banners that are used to cross-sell or promote other portions of your site. Essentially, you are measuring the click-thru rate of the banner, where you track the number of times the banner is displayed and divide that by the number of times the banner is clicked on. This rate is then monitored to watch to see when it just drops off. When you see this, you can recommend to your design team that this screen real estate be uses for a new promotion.


Page Views / Visit

One of the “out of the box” metrics that you see in a lot of web analytic tools is the average page views per visit. Many companies like to see this metric in their reports. It is the total number of pages viewed by your site visitors, divided by the number of visits to the site.



PPC simple means “pay per click”, which means that you get charged each time a person clicks on the link or ad to get to your website. Paid search is an example of PPC, as is sponsored links. Another term that is used along with PPC is “CPC”, which is the cost per click. This is how much the search engine charges for paid search results or sponsored links.


Referring Domain

If a person comes to your site, linked from another website, that site is known as the referring domain. When I look at referring domains, I like to look for things like coupon sites, deal sites, and other sites that I know nothing about that are bringing traffic to my site. Since you are not in control of these links, you can’t create campaign codes for these visits.


Natural Search

Natural search is also known as “organic” search. This is where your site gets listed in the search results on the main page, where you are not paying for placement.  The process of trying to optimize your website to rank high in natural search for various keywords is known as “SEO”, or search engine optimization. While most web analytic packages will show you visits from organic search, along with the keywords, you can’t tag an organic search visit with a campaign parameter, since the search engine is simply returning existing URLs of your site.


Direct / Bookmark

When a person comes to your site without the benefit of a referring domain or ad campaign, this is known as a direct visit. Since it is not possible to know if a person either typed in the URL or had it bookmarked in their browser, there is no distinction between the two. The percentage of these visits that are said to come from direct/bookmark is often under-reported, as many web users use the search bar to enter a website name, such as facebook.com. The first result is either a sponsored result or a natural search result for the site, and when the visitor clicks on the link, the analytics package treats this as a visit from paid search or organic search, and not a direct visit.


Cost per Click

This term is used in online marketing. Cost per click is also known as “CPC”. It is how much you pay, on average, each time someone clicks through from a link or ad to your website.


Cart Abandonment Rate

In e-commerce, products are often placed in an online shopping cart. If the prospect adds items to a cart, but never completes the sale, this is known as a cart abandonment. One of the jobs of a web analyst is to figure out why carts are being abandoned, and to make suggestions to reduce this rate. Some organizations also want to know the dollar value of abandoned carts to get an idea of how much money they are leaving on the table, theoretically.


Frequency of Purchase

Customers are often segmented by how often they make a purchase. This is known as “frequency of purchase”.


Average Order Size

The average order size (AOS) is the total revenue from online sales divided by the number of sales. Plenty of experimentation and testing is done on how to increase the AOS. Amazon does this by showing recommended additional products, or what customers tend to purchase together.


Self-Service Transactions

A self-service transaction is one where the user completes some action using the web instead of calling the help desk or using direct mail to complete the transaction. A few of my clients wanted to be able to measure call volume reductions as more of the customer needs are satisfied online. Self service transactions can include things like password resets, accessing reports that would typically be mailed, resolving problems using information from FAQs instead of picking up the phone and calling, obtaining a return authorization, or anything else where you want to encourage your site’s to use the site to conduct business instead of calling a person.


Media Site Metrics

You may hear your marketing team talk about media sites. These are typically news sites or portals where ads can be placed. Examples of media sites include CNN, Fox News, ESPN, Web MD, Yahoo, AOL, Huffington Post, and similar sites. Media sites typically make their money from ads on the site. Banner revenue is often based on the number of impressions the banner gets. Sponsored Links on pages earn revenue to the media site when a link is clicked. If you are the analyst for a media site, your managers may be interested in the number of page views, key content areas and what the key pages are where people visit.


Cross-Sell Opportunities

If you are in the financial services business (insurance, investments, banking, mortgages, etc…), one of the purposes of the site is to sell new products to existing clients. This is known as “cross sell”. One measure of success will be to know how often your site manages a cross sell. To do this, when a customer logs in, the backend logic of the site coding needs to be able to know what type of customer this is and what types of products the customer does not have. For example, if your company provides auto loan financing and mortgages, if a customer logs in to pay her bill, and this customer does not have a mortgage with the company, a banner can be displayed on the main page that promotes a refinance program or getting a quote for a home mortgage. When this page is loaded, a “cross sell opportunity” event would be set. If the visitor clicks through and takes a step toward getting a loan, such as filling out a mini-loan app or requesting information, then this would be considered as a cross-sell success.


Lead Generation

Many companies don’t transact business entirely on the web. For example, people typically don’t buy cars online, or things like sunrooms, health insurance, auto insurance, and other large-ticket products. Another example of this includes consulting agencies, accounting firms, law firms, financial planners, and so on. The website is typically used to generate a lead that then gets followed up on by a sales rep. So, measuring the number of leads generated is a key success metric, along with the campaigns that generated these leads. Ideally, you would either use campaign stacking capabilities in your web analytics tool to associate leads with the campaign string that brought the visitor to this point, and perhaps some sort of Lead ID that can be stored with your data, and assigned to the prospect. This way, if you have the right talent at your company, your company will then be able to measure the value that each of the campaign touches had in terms of revenue, and tie in the online behavior with offline revenue.


Registration Rate

This metric is often used to measure how well your site is doing in getting visitors to register for something. It can be a newsletter, email list  (for discount offers or special promotions), or anything else that gets the visitor one step close to being a prospect. Registration Rate is often used in B2B sites, where the purpose of the site is not necessarily to sell something to the visitor from the site.


Login Frequency

Another B2B related metric is the login frequency. It can be broken down to a series of frequency buckets, such as once or twice a month, three to ten times a month, 11 to 20 times, and so on. You would show what percentage of registered users login at each frequency. This will let you see how important your site is to your user base, with the goal of trying to move users from the lower-frequency buckets to the higher frequency buckets.


Download of Whitepapers

Many B2B sites offer something called “whitepapers” for download. A whitepaper is basically an in-depth report on some topic related to the business. For example, I have seen tag management system companies offer a whitepaper on how to measure the ROI of a TMS. A digital consulting company provided a whitepaper on how to use data appliances to simplify the “big data” problem facing many companies. The goal of putting out whitepapers is twofold. They can be given out without registration to anyone who wants it, with the goal of making a name for the company, or they can be offered only upon registration, with the goal of acquiring the prospect’s contact information for their marketing team. One mistake I often see is when a company requires registration in order to get the whitepaper, but simply provides the download link after registration, instead of sending it to their email address or making them verify the email by clicking on a link in the email to get to the download page. There have been many times where I have registered to get a whitepaper under the name of Mickey Mouse, with some bogus email address. If I can get the info without providing a valid email address, why should I? In any case, the downloading of a whitepaper should be considered as a success event and tracked accordingly.


Conversion Funnel

A conversion funnel is a key element of any lead generation or ecommerce site. It can either be page-based or event-based, and it shows the number of prospects who get to each stage of the conversion process. I am not going to go into detail on this one, as if you don’t know what a conversion funnel is, and you are a web analyst, you have bigger things to worry about! One key item to monitor in a conversion funnel is the dropout rate at each stage. If you can identify the problems that a prospect may be having at each step and make the necessary changes to improve the user experience, you can reduce the dropout rates and increase the overall conversion rate.


How to Measure ROI

As a web analyst, you win points (and perhaps a bonus and raise) by showing your company how to increase the ROI on your site, and how to measure that ROI. But, in order to do this, you need to understand what this all means. If you don’t already know, ROI stands for “Return on Investment”. It is the money that the company makes on money invested. Basically, it is the gain from an investment minus the cost of the investment, divided by the cost of the investment. Suppose for example, you have a baseline of an average of 10,000 orders per month from 434,000 visitors. That is a conversion rate of 2.30%. If your average revenue per sale is $50, your total revenue would be $500,000 from these visitors. If, through your analysis, you figure out how to increase the conversion rate to 3.1%, your resulting number of orders would be 13,454, for a revenue total of $672,700, or a difference of $172,700. If it cost your company $50,000 to make these improvements, your ROI would be ($172,000 – $50,000) / $50,000, or 245%. Note that this ROI was based only on the gross revenue, and does not factor in the cost of goods or services sold. In a true ROI model, you need to look at the net revenue after the cost of goods sold before being able to measure ROI. In the above example, suppose that on a $50 sale, your company makes $20. At a 2.3% conversion rate, the revenue on the 10,000 customers would be $200,000, instead of $500,000. At a 3.1% conversion rate, the 13,454 customers would provide a total of $269,080 in net revenue, for an ROI of 138%.


I hope this overview of some basic definitions will help you better understand what type of metrics to collect on your website and how to use these metrics to improve your site in meeting your organizations’ goals.



No comments yet — be the first.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: